A short written summary of what a public reel says, with three to five key points, built from its transcript. One call does both steps: the reel is transcribed if it has not been already, then a model writes the summary.
| parameter | type | description |
|---|---|---|
| url | string, required | A public Instagram post, reel or profile URL. Share links are resolved. |
| fresh | boolean | Skip the stored copy and read live, at the same price. |
| fields | string | Comma-separated field paths to keep. |
| format | json | JSON envelope. |
4 credits for the reel + 20 per started minute. Failed calls are never charged, and empty results are free up to 25 a day. Stored copies are stored once, served forever.
One object: a summary of about 150 words, an array of key points, the video's duration, and a language field that is null on Instagram today. The summary is stored with the transcript, so a repeat call for the same reel is served from storage.
Image posts, carousels and videos over ten minutes return 422 before anything is transcribed; private or deleted posts return 404. None of these are charged.
curl "https://api.instadata.dev/v1/instagram/summary?url=https%3A%2F%2Fwww.instagram.com%2Freel%2FC8xQvZ2sVAb%2F" \ -H "Authorization: Bearer ik_live_..."
{
"data": {
"summary": "A short pastry tutorial: cold butter is cut into flour, bound with ice water, rested, then rolled and folded twice before baking. …",
"key_points": [
"Keep the butter cold",
"Rest the dough before rolling",
"Fold twice for layers"
],
"language": null,
"duration_sec": 41
},
"request_id": "req_01J…"
} | field | type | note |
|---|---|---|
| summary | string | About 150 words describing what the video says. |
| key_points | string[] | Three to five short points, in the order they come up. |
| language | null | Always null on Instagram today: no language is detected upstream. |
| duration_sec | number|null | Length of the video in seconds. Billing counts started minutes of this value. |
4 credits for the summary plus 20 per started minute of video, the transcription's upstream price. A reel under a minute is 24 credits. If the transcript already exists in storage from an earlier call, the per-minute part is still charged: the price is per result, not per upstream run.
The summary is written by a language model from the transcript. It is good at the gist and can be wrong on specifics, numbers and names. Quote the transcript, not the summary.
A music-only reel still costs one started minute and yields a summary of whatever the transcription heard, which can be noise. Image posts and carousels return 422 before anything is transcribed, uncharged.
4 credits plus 20 per started minute of video. A reel under a minute is 24 credits. Failed calls (no video, over ten minutes, private, deleted) cost nothing.
No. The summary endpoint runs the transcript step itself when the reel has not been transcribed yet, and reuses the stored transcript when it has.
About 150 words, plus three to five key points. Add fields=summary to receive only the paragraph.
The language spoken in the reel. The language field is null on Instagram today, so read it from the text itself.
Yes, alongside the transcript. A repeat call for the same reel returns from storage at the same price.